Dockerfile

FROM node:24-slim

# git is needed for HF Dataset DB sync
RUN apt-get update && apt-get install -y git poppler-utils \
    && rm -rf /var/lib/apt/lists/*

# Install pnpm
RUN npm install -g pnpm@10

WORKDIR /app

# ── Copy workspace manifests first (layer-cache friendly) ──────────────────
COPY package.json pnpm-workspace.yaml pnpm-lock.yaml ./

COPY artifacts/api-server/package.json ./artifacts/api-server/
COPY artifacts/raqim/package.json      ./artifacts/raqim/

COPY lib/db/package.json                             ./lib/db/
COPY lib/api-client-react/package.json               ./lib/api-client-react/
COPY lib/api-spec/package.json                       ./lib/api-spec/
COPY lib/api-zod/package.json                        ./lib/api-zod/
COPY lib/integrations-openai-ai-server/package.json  ./lib/integrations-openai-ai-server/
COPY lib/integrations-openai-ai-react/package.json   ./lib/integrations-openai-ai-react/
COPY scripts/package.json                            ./scripts/

# ── Install all dependencies ───────────────────────────────────────────────
RUN pnpm install --frozen-lockfile

# ── Copy full source ───────────────────────────────────────────────────────
COPY . .

# ── Build frontend (BASE_PATH=/ for single-origin serving) ────────────────
RUN BASE_PATH=/ PORT=3000 NODE_ENV=production \
    pnpm --filter @workspace/raqim run build

# ── Build backend ──────────────────────────────────────────────────────────
RUN pnpm --filter @workspace/api-server run build

# ── Persistent data dirs ───────────────────────────────────────────────────
# /data is the HF Spaces persistent volume; /data/uploads stays across restarts.
# /tmp/uploads is NOT used in production — /data/uploads is used instead.
RUN mkdir -p /data /data/uploads

# ── Git identity for HF Dataset sync ──────────────────────────────────────
RUN git config --global user.email "sync@raqim.app" && \
    git config --global user.name "RAQIM Sync"

# ── Runtime config ────────────────────────────────────────────────────────
ENV PORT=7860
ENV NODE_ENV=production
ENV FRONTEND_DIST=/app/artifacts/raqim/dist/public
ENV DB_PATH=/data/raqim.db
ENV MIGRATIONS_DIR=/app/lib/db/drizzle

EXPOSE 7860

# Startup sequence:
# 1. Pull DB backup from HF Dataset ONLY if /data/raqim.db is absent (first boot)
# 2. Seed admin user (skips if already exists)
# 3. Background loop pushes DB every 60 s (was 10 min) to minimise data loss
# 4. Run server as PID 1 via exec (exit code propagates to container)
# 5. On exit, push DB one final time
CMD ["sh", "-c", "\
  mkdir -p /data/uploads /data/tessdata ; \
  node /app/scripts/db-hf-sync.mjs pull || true ; \
  node /app/scripts/seed-admin.mjs || true ; \
  trap 'node /app/scripts/db-hf-sync.mjs push || true' EXIT ; \
  (while true; do sleep 60 && node /app/scripts/db-hf-sync.mjs push || true; done) & \
  exec node --enable-source-maps /app/artifacts/api-server/dist/index.mjs \
"]